8/13/2019 161142-1

1/85

Technical Report Documentation Page1. Report No.SWUTC/11/161142-1

2. Government Accession No. 3. Recipient's Catalog No.

4. Title and SubtitleEvaluation of Mobile Source Greenhouse Gas Emissionsfor Assessment of Traffic Management Strategies

5. Report DateAugust 2011

6. Performing Organization Code

7. Author(s)

Qinyi Shi and Lei Yu8. Performing Organization Report No.

Report 161142-19. Performing Organization Name and AddressTexas Southern University3100 Cleburne AvenueHouston, TX 77004

10. Work Unit No. (TRAIS)

11. Contract or Grant No.10727

12. Sponsoring Agency Name and AddressSouthwest Region University Transportation CenterTexas Transportation InstituteTexas A&M University SystemCollege Station, Texas 77843-3135

13. Type of Report and Period CoveredResearch ReportSeptember 2009 August 201114. Sponsoring Agency Code

15. Supplementary NotesSupported by general revenues from the State of Texas.

16. Abstract

In recent years, there has been an increasing interest in investigating the air quality benefits of trafficmanagement strategies in light of challenges associated with the global warming and climate change.However, there has been a lack of systematic effort to study the impact of a specific traffic managementstrategy on mobile source Greenhouse Gas (GHG) emissions. This research is intended to evaluatemobile source GHG emissions for traffic management strategies, in which a Portable EmissionMeasurement System (PEMS) is used to collect the vehicles real-world emission and activity data, and

a Vehicle Specific Power (VSP) based modeling approach is used as the basis for emission estimation.Three traffic management strategies are selected in this research, including High Occupancy Vehicle(HOV) lane, traffic signal coordination plan, and Electronic Toll Collection (ETC). In the HOV lanescenario, CO 2 emission factors produced by the testing vehicle using HOV lane and the correspondingmixed flow lane are compared. In the evaluation of traffic signal coordination, total CO 2 emissions

produced under the existing coordinated signal timing and the emulated non-coordinated signal timingalong the same designed testing route are compared. In the study about ETC, total CO 2 emissions

produced by the testing vehicle around an ETC station and a Manual Toll Collection (MTC) stationlocated on the same toll road segment are estimated and compared. The results demonstrated that HOVlane, well-coordinated signal timing, and ETC are all effective measures to reduce mobile source GHGemissions, although the level of effectiveness is shown to be different for different strategies.

17. Key WordsGreenhouse Gas (GHG) Emissions, VehicleSpecific Power (VSP), Traffic ManagementStrategy, Evaluation Method, Portable EmissionMeasurement System (PEMS)

18. Distribution Statement No restrictions. This document is available to the public through NTIS: National Technical Information Service5285 Port Royal RoadSpringfield, Virginia 22161

19. Security Classify (of this report)Unclassified

20. Security Classify (of this page)Unclassified

21. No. of Pages85

22. Price

Form DOT F 1700.7 (8-72) Reproduction of completed page authorized

8/13/2019 161142-1

2/85

ii

8/13/2019 161142-1

3/85

8/13/2019 161142-1

4/85

8/13/2019 161142-1

5/85

v

ABSTRACT

In recent years, there has been an increasing interest in investigating the air quality benefits of

traffic management strategies in light of challenges associated with the global warming and

climate change. However, there has been a lack of systematic effort to study the impact of a

specific traffic management strategy on mobile source Greenhouse Gas (GHG) emissions. This

research is intended to evaluate mobile source GHG emissions for traffic management strategies,

in which a Portable Emission Measurement System (PEMS) is used to collect the vehicles real-

world emission and activity data, and a Vehicle Specific Power (VSP) based modeling approach

is used as the basis for emission estimation. Three traffic management strategies are selected in

this research, including High Occupancy Vehicle (HOV) lane, traffic signal coordination plan,and Electronic Toll Collection (ETC). In the HOV lane scenario, CO 2 emission factors produced

by the testing vehicle using HOV lane and the corresponding mixed flow lane are compared. In

the evaluation of traffic signal coordination, total CO 2 emissions produced under the existing

coordinated signal timing and the emulated non-coordinated signal timing along the same

designed testing route are compared. In the study about ETC, total CO 2 emissions produced by

the testing vehicle around an ETC station and a Manual Toll Collection (MTC) station located on

the same toll road segment are estimated and compared. The results demonstrated that HOV lane,

well-coordinated signal timing, and ETC are all effective measures to reduce mobile source

GHG emissions, although the level of effectiveness is shown to be different for different

strategies.

8/13/2019 161142-1

6/85

vi

8/13/2019 161142-1

7/85

vii

EXECUTIVE SUMMARY

Issues regarding Greenhouse Gas (GHG) emissions have attracted world-wide attention. In the

transportation sector, emissions from on-road vehicles are known as a major source of GHG

emissions. As we know, the implementation of different traffic management strategies will

result in changes in emission levels for different emission species, therefore, these strategies can

potentially be very effective approaches to reduce mobile source GHG emissions, especially a

vehicles CO 2 emissions. However, due to real-world data constraints and limitations associated

with current mobile source emission models, there has been a lack of systematic effort to study

the impact of a specific traffic management strategy on mobile source GHG emission control. In

this context, the primary objectives of this research are to: (1) develop an emission estimationmethodology to quantify a vehicles CO 2 emissions in a real-world traffic network; (2) design

field testing scenarios to collect a vehicles real-world emission and operational data with versus

without the implementation of the selected traffic management strategies; and (3) provide a

quantitative evaluation of the selected traffic management strategies in terms of their

effectiveness on reducing a vehicles CO 2 emissions.

In this research, the evaluation of traffic management strategies is fulfilled by a combined use of

the field data collected by a Portable Emission Measurement System (PEMS) and a vehicle

specific power (VSP) based modeling approach. The general methodology includes

comprehensive data collection, the application of state-of-the-art vehicle emission modeling

approach, and a thorough evaluation of the selected traffic management strategies.

Two parts of real-world data are needed to perform the proposed evaluation study. One part

includes the data collected for the purpose of developing the modeling approach that meets

specific needs of the emission calculation in this study; and the other part includes the datacollected in the designed testing areas to facilitate the case-specific traffic management

assessment. A VSP-based emission modeling approach is developed to quantify a vehicles CO 2

emissions during its regular operations. The basic methodology for this modeling approach is

binning second-by-second VSP data and computing the average emission rate in each bin. With

8/13/2019 161142-1

8/85

viii

this partition, the average emission rate of a particular type of pollutant in that bin for a specific

vehicle can be calculated. The evaluation approach is based on the comparison of emissions

produced with versus without the implementation of a specific traffic management strategy.

Since this research focuses on existing traffic management strategies, the testing vehicles real-

world emissions and operational data can be directly collected using the PEMS equipment. In

the meantime, case-specific data collection plans need to be developed so that the vehicles

emissions under the scenario without the implementation of the selected traffic management

strategy can be calculated in the real-world setting. The VSP-based emission modeling approach

makes it possible to perform emission calculations by needing only the vehicles speed and

acceleration, therefore, in this study, a vehicle equipped with a GPS device is used to run with

the vehicle equipped with the PEMS unit in a synchronized way under different scenarios. In

this way, a pair of paralleled datasets can be obtained for the purpose of comparison.

The following conclusions are drawn from this research:

First, PEMS represents an advanced emission data collection technology. Its ability of collecting

a vehicles second-by-second emission and activity data during its regular operations provides

significant advantages over all the traditional emission measurement methods. It can be applied

not only in transportation related air quality modeling and analysis, but also in the assessment of

traffic management strategies.

Second, the proposed emission estimation methodology is a combination of the advantages of

field testing approach and the latest modeling approach. The experimental design of the field

testing scenarios provides a pilot study on the impact of a specific traffic management strategy

on mobile source GHG emissions using data collected in the real-world traffic network. The

VSP-based emission modeling approach provides a credible basis for emission estimation. The

validation results indicate that the accuracy rate of the proposed modeling approach is about

90%.

8/13/2019 161142-1

9/85

ix

Third, HOV lane, well-coordinated signal timing, and ETC are all effective measures to reduce

mobile source GHG emissions; however, the level of effectiveness is different for different

strategies.

Fourth, the results from HOV lane analysis illustrate that the testing vehicle produces less mass

CO2 emissions per mile by using HOV lane during peak periods. Without the consideration of

the effect of HOV lane on vehicle miles traveled, the emission reduction rate on the first testing

day is 3.56 percent, and due to an increased traffic demand on the corresponding MF lane on the

second testing day, the emission reduction rate by using HOV lane increased to 10.42 percent.

Fifth, based on the comparison of CO 2 emissions generated by the testing vehicle under the

existing coordinated signal timing and those generated under the emulated non-coordinatedsignal timing, it is found that the non-coordinated signal timing designed in this study leads to

about 56 percent increase in CO 2 emissions. It is also found that the increase of traffic flow may

compromise the effectiveness of signal coordination in terms of their influence on mobile source

GHG emission control.

Finally, the results from the ETC analysis shows that the total CO 2 emissions produced by the

vehicle around the ETC station are only 30 percent of those produced around the corresponding

MTC station; therefore, ETC is a very effective traffic management strategy for reducing mobile

source GHG emissions.

In order to fulfill a more comprehensive analysis about the relationship between traffic

management strategies and mobile source GHG emissions, the following recommendations are

made for future study:

1. Improve the VSP-based emission modeling approach by increasing the size of the

database used for the model development and develop a finer VSP binning method.

2. Evaluate the impact of traffic management strategies on mobile source GHG

emissions with the use of different vehicle types.

8/13/2019 161142-1

10/85

x

3. Incorporate cost-benefit analysis into the evaluation of traffic management strategies,

such as construction cost, operation and maintenance cost, and comprehensive air

quality benefits.

4. Evaluate the impact of traffic management strategies on regional mobile source GHG

emission reduction.

8/13/2019 161142-1

11/85

xi

TABLE OF CONTENTS

ABSTRACT .................................................................................................................................. V

EXECUTIVE SUMMARY ....................................................................................................... VII

TABLE OF CONTENTS ........................................................................................................... XI

LIST OF FIGURES ................................................................................................................. XIV

LIST OF TABLES ..................................................................................................................... XV

DISCLAIMER ......................................................................................................................... XVI

ACKNOWLEDGMENT ........................................................................................................ XVII

CHAPTER 1: INTRODUCTION ................................................................................................ 1

1.1 Background of Research .................................................................................................................................... 2

1.1.1 Mobile Source GHG Emission ....................................................................................................................... 2

1.1.2 Mobile Source GHG Emission Reduction Practice ....................................................................................... 3

1.1.3 Mobile Source GHG Emission Estimation ..................................................................................................... 4 1.1.4 Issues and Research Gaps ............................................................................................................................. 5

1.2 Objectives of Research ....................................................................................................................................... 6

1.3 Outline of the Report .......................................................................................................................................... 7

CHAPTER 2: LITERATURE REVIEW .................................................................................... 9

2.1 Major GHG Emission Measurement Methods ................................................................................................... 9

2.1.1 Chassis Dynamometer Testing ........................................................................................................................ 9

2.1.2 Tunnel Study .................................................................................................................................................. 10

2.1.3 Remote Emission Sensing (RES) ................................................................................................................... 10

2.1.4 Portable Emission Measurement System (PEMS) ......................................................................................... 11

2.2 Mobile Source GHG Emission Models ............................................................................................................ 12

2.2.1 MOBILE6 Model ........................................................................................................................................... 12

2.2.2 NONROAD Model ......................................................................................................................................... 13

2.2.3 National Mobile Inventory Model (NMIM) ................................................................................................... 13

8/13/2019 161142-1

12/85

8/13/2019 161142-1

13/85

xiii

4.4 Evaluation and Analysis of Electronic Toll Collection Scenario ..................................................................... 46

CHAPTER 5: SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS ................... 50

5.1 Summary of the Study ...................................................................................................................................... 51

5.2 Conclusions ...................................................................................................................................................... 51 5.3 Recommendations ............................................................................................................................................ 53

APPENDIX................................................................................................................................... 55

APPENDIX A: ACRONYMS AND ABBREVIATIONS ..................................................................................... 55

APPENDIX B: EMISSION DATABASE STRUCTURE AND EXPLANATION .......... ........... ........... ........... ... 58

REFERENCES ............................................................................................................................ 61

8/13/2019 161142-1

14/85

xiv

LIST OF FIGURES

Figure # Page

FIGURE 1 SELECTED PICTURES DURING AN AXION -BASED EMISSION TEST ................................................................. 24

FIGURE 2 GEOLOG GPS DEVICE USED IN THE TEST .................................................................................................... 25

FIGURE 3 STUDY AREA FOR HOV LANE A NALYSIS ..................................................................................................... 30

FIGURE 4 3-M ONTH AVERAGE SPEED ON THE SELECTED FREEWAY SEGMENT ON I-45 N .......... .......... ........... .......... .. 31

FIGURE 5 STUDY AREA FOR SIGNAL COORDINATION A NALYSIS ................................................................................. 32

FIGURE 6 STUDY AREA FOR ETC A NALYSIS ............................................................................................................... 33

FIGURE 7 AVERAGE CO2 EMISSION R ATE AND DATA FREQUENCY FOR EACH VSP BIN .............................................. 36

FIGURE 8 COMPARISONS OF VSP DISTRIBUTIONS ON HOV LANE AND MF LANE ....................................................... 40 FIGURE 9 COMPARISON OF CO2 EMISSIONS FOLLOWING EXISTING COORDINATED SIGNAL TIMING AND EMULATED

NON-COORDINATED SIGNAL TIMING IN TOTAL ................................................................................................. 42

FIGURE 10 COMPARISON OF CO2 EMISSIONS FOLLOWING EXISTING COORDINATED SIGNAL TIMING AND EMULATED

NON-COORDINATED SIGNAL TIMING BY CYCLE ................................................................................................ 43

FIGURE 11 COMPARISON OF VSP DISTRIBUTIONS IN EACH CYCLE ............................................................................. 45

FIGURE 12 PROCEDURE OF DATA SELECTION AROUND MTC STATION AND ETC STATION ......................................... 47

FIGURE 13 COMPARISON OF TOTAL CO2 EMISSIONS PRODUCED AROUND ETC STATION AND MTC STATION ............ 48

FIGURE 14 COMAPRISON OF CO2 EMISSIONS EACH TIME THE VEHICLE PASSES ETC STATION AND MTC STATION ... 49

8/13/2019 161142-1

15/85

xv

LIST OF TABLES

Table # Page

TABLE 1 DEFINITION OF MOVES OPERATING MODE ATTRIBUTES FOR R UNNING E NERGY CONSUMPTION ................ 28

TABLE 2 I NFORMATION ABOUT THE TESTING VEHICLE FOR MODEL DEVELOPMENT ................................................... 35

TABLE 3 AVERAGE CO2 EMISSION R ATE FOR EACH BIN ............................................................................................. 37

TABLE 4 COMPARISON OF TOTAL CO2 EMISSIONS BASED ON MODELING APPROACH AND PEMS EMISSION DATA ... 38

TABLE 5 COMPARISON OF CO2 EMISSION FACTORS USING HOV LANE AND MF LANE .............................................. 39

TABLE 6 I NFORMATION ABOUT THE TESTING VEHICLE FOR EVALUATION ON ETC ..................................................... 46

8/13/2019 161142-1

16/85

xvi

DISCLAIMER

Contents of this report reflect the views of the authors who are responsible for the facts and the

accuracy of the information presented herein. This document is disseminated under the

sponsorship of the U.S. Department of Transportation University Transportation Centers

Program, in the interest of information exchange. The U.S. Government assumes no liability for

the contents or use thereof.

8/13/2019 161142-1

17/85

xvii

ACKNOWLEDGMENT

This publication was developed as part of the University Transportation Center Program which is

funded, in part, with general revenue funds from the State of Texas.

8/13/2019 161142-1

18/85

xviii

8/13/2019 161142-1

19/85

1

CHAPTER 1: INTRODUCTION

Climate change is one of the most serious worldwide environmental problems. Most scientists

agree that the major cause of climate change is greenhouse gases (GHGs) resulting from human

activities. In the United States, transportation is a major source as well as the fastest growing

sector of GHG emissions. In addition, almost all of the increases in transportation related GHG

emissions since 1990 are brought about by on-road vehicles, i.e. mobile source, in the form of

carbon dioxide (CO 2) (EPA, 2010a; EIA, 2008). Therefore, if we do not promptly and

substantially reduce mobile source GHG emissions, adverse consequences of the global warming

may only become worse in number and intensity.

It is well understood that the implementation of different traffic management strategies will

result in changes in emission levels for different emission species; thus, traffic management

strategies are potentially a very effective approach to reduce different types of mobile source

emissions. However, due to the real-world data constraints, limitations associated with current

GHG emission models, and a lack of comprehensive analysis of traffic management related

GHG emission reduction methodologies, there has been a lack of systematic attempt to study the

impact of a specific traffic management strategy on mobile source GHG emissions.

To address the gap that exists in the current practice, this research intends to provide a

quantitative evaluation of three selected traffic management strategies in terms of their

effectiveness on reducing a vehicles CO 2 emissions with a combined use of field testing

approach and modeling approach. The merit of the proposed methodology is that it not only

uses the existing model as the basis for emission estimation, but also combines it with real-world

tests. Using the proposed methodology, a vehicle CO 2 emissions for scenarios with versuswithout the implementation of the selected traffic management strategies are estimated and

compared.

8/13/2019 161142-1

20/85

2

1.1 Background of Research

1.1.1 Mobile Source GHG Emissions

GHGs mainly include CO 2, methane (CH 4), nitrous oxide (N 2O), and hydrofluorocarbons

(HFCs), in which CO 2 from fossil fuel combustion has accounted for approximately 80 percent

of global warming potential (GWP) weighted emission in 2007 (EPA, 2009a). In the United

States, the transportation sector accounts for about 33 percent of total CO 2 emissions, giving the

largest share of any end-use economic sector. Nearly 60 percent of transportation-related CO 2

emissions result from gasoline consumption for personal vehicle use. The remaining come from

other transportation activities, such as the combustion of diesel fuel in heavy-duty vehicles and

jet fuel in aircrafts (EPA, 2010a). Therefore, the mobile source GHG emissions discussed in this

study focus on vehicles CO 2 emissions.

Four key factors affect mobile source GHG emissions, including vehicle technology, fuel

economy, vehicle miles traveled (VMT), and vehicle/system operations (FHWA, 2008). The

traffic management has been implemented to reduce the growth in VMT and improve the

efficiency of transpiration system operations. Related policy scenarios include congestion

pricing, speed limit reduction, public transit development, etc. (Rodier, 2008). Transportationengineers and planners also look to switch to alternative fuels, use more fuel efficient vehicles,

and enhance Inspection and Maintenance (I/M) program for in-using vehicles to lower mobile

source GHG emissions.

The development of mobile source GHG emission inventory starts with an estimation of fuel

consumption from the Energy Information Administration (EIA) of the U. S. Department of

Energy. The fuel consumption statistics made by EIA are believed to be very accurate in

accounting for the combined fuel consumption of all economic sectors, but great uncertainty may

occur when one apportions the fuel- and sector- specific estimates to certain sources (Davis, et

al., 2007). The estimation of mobile source GHG emissions at the project level is usually

realized through the field-testing approach or modeling approach. Therefore, the accuracy and

8/13/2019 161142-1

21/85

3

applicability of these approaches are greatly affected by the selected GHG emission

measurement technologies or emission models.

1.1.2 Mobile Source GHG Emission Reduction Practice

Mobile source GHG emission reduction practices are conducted at different levels from federal

government to local transportation agencies. U.S Environmental Protection Agency (EPA) plays

a significant role in this area (EPA, 2010b). EPAs mobile source GHG emission reduction

programs include Clean Energy-Environment State Partnership, Climate Leaders, Energy Star,

and EPA Office of Transportation and Air Quality Voluntary Programs, such as National Clean

Diesel Campaign (NCDC), SmartWay Transport Partnership, Clean School Bus USA, Best

Workplaces for Commuters, and EcoCar (EPA, 2010c). Other federal mobile source GHG

emissions reduction initiatives include Climate VISION Partnership, Tax Incentives to Reduce

GHG Emissions, and Voluntary Greenhouse Gas Reporting Program. These programs promote

voluntary GHG reduction by encouraging auto manufacturers to implement cost-effective clean

energy and environmental strategies, developing comprehensive GHG emission control

regulations, and providing targeted incentives to spur the use of more energy-efficient

technologies.

At the state level, specific climate policies are adopted to address the problem of mobile source

GHG emissions. Even though, currently, there is no federal requirement to report GHG

emissions, as of November 2008, 18 states have imposed mandatory reporting of GHG emissions

(EHS Today, 2009). California has passed legislation requiring the state's Air Resources Board

to set GHG emission standards for new passenger cars and light-duty trucks from the model year

2009 and later. In addition, The California Zero-Emission Vehicle Incentive Program provides

grants of up to $9, 000 per vehicle toward the purchase or lease of new zero-emission vehicles.

The Advanced Travel Center Electrification (ATE) program is conducted in the states of

Arkansas, Georgia, New York, and Tennessee, which provides energy-efficient heating,

ventilation, and cooling systems (HVAC) for use by truckers at travel centers and other areas

where drivers stop and idle their vehicles.

8/13/2019 161142-1

22/85

4

At local and project level, various practices have also been pursued to reduce mobile source

GHG emissions. Such practices include improving traffic management and public transportation

amenities to optimize transportation system operation; accelerating vehicle retirement to reduce

transportation fuel consumption; and implementing financial incentives, pricing regulations, car

sharing, and broader use of telecommunication technologies to reduce travel and congestion

(Euritt et al., 1996; Greene and Schafer, 2003).

1.1.3 Mobile Source GHG Emission Estimation

Accurate quantification of vehicles emissions is the basis for evaluating the air quality benefits

of any traffic management strategies. Since 1970s, various emission measurement methods have

been developed to collect vehicles emission data during field testing, in which chassis

dynamometer testing, tunnel testing, remote sensing, and Portable Emission Measurement

Systems (PEMS) are the most widely used methods. Even though these technologies are mainly

developed to measure hazardous pollutants from vehicle exhausts, such as hydrocarbons (HC),

nitrogen oxide (NO X), and carbon monoxide (CO), they all have the ability to quantify CO 2

emissions from on-road vehicles.

A wide range of GHG emission models have also been developed to estimate the amount ofmobile source GHG emissions produced under various traffic management strategies. Direct

GHG emission estimation models include MOBILE6 model, NONROAD model, National

Mobile Inventory Model (NMIM), EMission FACtors Model (EMFAC), and Climate Leadership

in Parks (CLIP) model (EPA, 2006a; EPA, 2006b; EPA, 2006c; CARB, 2010; NPS and EPA,

2007). These models focus on transportation sources, and are designed to develop emission

factors or emission estimates for pollutants emitted during vehicles use. Greenhouse Gases,

Regulated Emissions, and Energy use in Transportation (GREET) Model, Lifecycle Emissions

Model (LEM), and Motor Vehicle Emissions Simulator (MOVES) are also capable of GHG

emissions quantification. These models are termed as life-cycle models, which take into

consideration not only tailpipe emissions but also pre-combustion emissions (ANL, 2009;

Delucchi, 2002; EPA, 2010d).

8/13/2019 161142-1

23/85

5

Of all the GHG emission models examined, EPAs MOVES provides the most functionality and

applicability for conducting different types of transportation GHG analysis (ICF Consulting,

2006). MOVES uses Vehicle Specific Power (VSP) to characterize emission rates for the

running exhaust emission process, which combines into one single parameter numerous physical

factors that are influential to vehicle fuel consumption and emissions, such as vehicle speed,

acceleration, road grade, and road load parameters (Koupal et al, 2002). In addition, existing

studies have found that the VSP binning approach has the most consistent performance when

matching VSP distribution with real-world CO 2 emissions per unit time (EPA, 2002). Therefore,

a VSP-based modeling approach is used as the basis for emission estimations in this research.

1.1.4

Issues and Research Gaps

Although a lot of research has been conducted to evaluate air quality benefits of specific traffic

management strategies, such as high occupancy vehicle (HOV) lane, bus exclusive lane, traffic

signal control plan, electronic toll collection (ETC), speed restriction, banning heavy duty

vehicles, and adaptive cruise control, current practices still face several barriers in further

integrating GHG emissions reduction in transportation planning. Particular challenges arise from

real-world data constraints, limitations associated with current GHG emission models, and lack

of systematic analysis of traffic management related GHG emission reduction methodologies.

Real-world data constraint is resulted from the limitation of traditional emission measurement

technologies. For instance, dynamometer testing takes place in optimum ambient conditions

(fixed temperature, pressure, and humidity) on a predefined driving cycle, which is unable to

reflect the vehicles emissions in real-world driving conditions; tunnel test does happen in the

real-world driving settings, but the results only represent average emission factors generated in

the tunnel; RES is good at giving instantaneous estimate of emission performance of a large

amount of vehicles in different classes, but which cannot capture vehicles corresponding

emission rates under different driving patterns (acceleration, deceleration, and cruising) and

engine temperature over an extended period of time.

8/13/2019 161142-1

24/85

6

Limitations associated with the existing GHG emission models are another barrier to incorporate

GHG emission control into traffic management. Take EPAs MOBILE6 as an example.

MOBILE6 model can perform CO 2 emission estimate, but the resulting CO 2 emission factors do

not vary with the vehicles speed or driving cycle. For this reason, MOBILE6 is inappropriate

for any kind of detailed traffic management planning or project level emissions analysis, which

is likely to involve changes of congestion levels and vehicle speeds (Grant et al, 2008).

Californias EMFAC model also estimates vehicle emissions by average trip speed. In addition,

EMFAC does not include speed corrections for most vehicle classes for CO 2. This model is

therefore insensitive to the impact of traffic conditions on the CO 2 emissions of vehicles in

different classes (CARB, 2007).

Because there are no regulations addressing GHG emissions from transportation sources (exceptCalifornia), the state department of transportation (DOT), metropolitan planning organizations

(MPOs), and other transportation agencies have limited experiences in analyzing the impact of

traffic management strategies on mobile source GHG emissions. In addition, researchers and

practitioners are concerned that the current modeling approaches that facilitate the GHG impacts

assessment are insufficient for being used to conduct the types of analysis necessary to

strategically address GHG emissions at project, local, and regional levels. Therefore, there is a

lack of systematic analysis of traffic management related GHG emission reduction

methodologies.

1.2 Objectives of Research

The research conducted in this study is motivated by the need to reduce mobile source GHG

emissions from the perspectives of transportation planning and traffic management. In light of

the above discussion, this research is intended to achieve three objectives:

1. Develop an emission estimation methodology to quantify a vehicles CO 2 emissions in a

real-world traffic network;

8/13/2019 161142-1

25/85

7

2. Design field testing scenarios to collect a vehicles real-world emission and operational data

with versus without the implementation of the selected traffic management strategies; and

3. Provide a quantitative evaluation of the selected traffic management strategies in terms of

their effectiveness on reducing a vehicles CO 2 emissions.

1.3 Outline of the Report

This report is organized into five chapters:

The first chapter provides readers with the background of mobile source GHG emissions, state-

of-the-art mobile source GHG emissions reduction practices, and state-of-the-practice mobile

source GHG emissions estimation. It also presents the problems identified in the current studies

on the evaluation of traffic management strategies for mobile source GHG emissions, which is

then followed by a description of research objectives.

The second chapter summarizes existing studies related to GHG emissions data collection

methods, mobile source GHG emission models, and state-of-the-art assessment of traffic

management strategies in terms of their impact on mobile source GHG emissions.

The third chapter describes the design of the study. It presents when, where, and how the CO 2

emissions produced by the testing vehicle are collected for the development of the proposed

emission modeling approach and for the evaluation of the selected traffic management strategies.

It introduces the methodology on how the VSP-based GHG emission modeling approach is

developed with a combined use of real-world data and state-of-the-art emission modeling

approach. The design of the assessment of traffic management strategies in terms of their impact

on mobile source GHG emission reduction is also introduced.

The fourth chapter presents the details of the results from this research, including the

establishment and validation of the proposed emission modeling approach, and a discussion

about the results generated from the application of this approach. An evaluation of the selected

traffic management strategies is also presented based on the resulting emission estimations.

8/13/2019 161142-1

26/85

8/13/2019 161142-1

27/85

9

CHAPTER 2: LITERATURE REVIEW

As one of principle causes to the air quality problems associated with the global warming andclimate change, mobile source GHG emissions have been given significant attention and the

research in this area has gone through a remarkable development. In order to gain better insights

into this field and conduct the research in this study in the most up-to-date setting, a

comprehensive literature review is of significant importance. This chapter reviews major GHG

emission measurement methods, state-of-the-art mobile source GHG emission models, and

existing studies on air quality benefit assessment of traffic management strategies. Based on the

review results, limitations that exist in the current research are presented and the methodology

that will be used in this research is identified.

2.1 Major GHG Emission Measurement Methods

The quantification of vehicles emissions depends on emission measurement methods. The

research in this study requires the use of emission measurement technology to collect mobile

source GHG emissions under different traffic management strategies. Currently, there are four

major emission measurement methods: Chassis Dynamometer Test, Tunnel Test, Remote

Sensing, and PEMS.

2.1.1 Chassis Dynamometer Testing

Chassis dynamometer is a standard tool for vehicle emission tests. It is designed to simulate the

road load to measure exhaust emissions via predefined driving schedules in an exhaust emission

laboratory (EPA, 2010e). The test system usually includes complex emissions sampling

equipment and exhausts emissions analyzer, which enable the system to measure exhaust

emissions based on the level of dynamics of each driving situation and specific characteristics of

each vehicle, such as model, year, mileage, engine type, fuel type, emission control standards,

etc. In recent years, chassis dynamometer has been improved by incorporating advanced roller

8/13/2019 161142-1

28/85

8/13/2019 161142-1

29/85

11

If the level of emission rate is above a certain threshold, a freeze-frame video system will then be

employed to digitize an image of the license plate number of the offending vehicle (Virginia

DOEQ, 2003; EPA, 2004b). This technology is able to accomplish an emission test at a specific

location during vehicles real-world operating conditions and generate a large amount of

emission data for different vehicle types within a short time and at a relatively low cost.

However, the major disadvantage of this technology is that it only gives an instantaneous

estimation of emission concentration, which means that the mass of emissions cannot be

obtained and the variation of emission rates under different driving modes and engine

temperatures cannot be reflected. In order to evaluate the impact of traffic management

strategies on mobile source GHG emissions, it is necessary to compare a vehicles mass

emissions for scenarios with versus without the implementation of a specific traffic management

strategy in an area over an extended period of time. Therefore, RES technology is not directlyapplicable for this study.

2.1.4 Portable Emission Measurement System (PEMS)

PEMS was developed for emission inventory and regulatory applications in late 1990s under the

lead of EPA. This measurement technology overcomes the limitations of chassis dynamometer,

tunnel test, and RES by being able to measure emissions during the actual use of vehicles in theirregular operations. It can collect the emission data on different road types, during different time

periods, and for various types of vehicles in an easy and convenient manner. At present, one of

the most advanced and representative PEMS products is OEM-2100 system developed by Clean

Air Technologies International, Inc. (CATI). This system has been verified by EPAs

Environmental Technology Verification (ETV) Program in 2003 for its precision and accuracy

(Myers, Kelly, Dindal, Willenberg, and Riggs, 2003). The latest version is OEM-2100AX Axion.

The Axion measures engine data using a set of sensors and reports engine and vehicle parameters

from an engine control unit (ECU) interface. This unit is capable of collecting gaseous variables,

including CO, CO 2, HC, NO X, O2, particulate matter (PM), and fuel consumption on a second-

by-second basis and supports real time text and graphic display. In addition, a global positioning

system (GPS) is included to provide the information about the testing vehicles location and

8/13/2019 161142-1

30/85

8/13/2019 161142-1

31/85

8/13/2019 161142-1

32/85

8/13/2019 161142-1

33/85

8/13/2019 161142-1

34/85

16

vehicle operating characteristics on emissions, therefore cannot be readily used for project-level

analyses.

2.3 Air Quality Benefit Assessment of Traffic Management Strategies

With the aid of available GHG emissions measurement technologies and various GHG emission

models, extensive research efforts have been made to assess the impact of traffic management

strategies on GHG emissions. Major traffic management and control measures that have been

investigated in terms of their impact on vehicles CO 2 emissions include HOV lane, bus

exclusive lane, traffic signal coordination plan, and ETC.

2.3.1 HOV Lane

Krimmer and Venigalla conducted an experimental study in the metropolitan Washington D.C.

area to compare the testing vehicles emissions on HOV lanes and mixed-flow (MF) lanes

(Krimmer and Venigalla, 2006). In their research, several hundred miles of on-road emissions

data were collected using PEMS by running pairs of nearly identical instrumented vehicles

simultaneously on the HOV and MF facilities. A major finding was that higher speeds in HOV

lanes resulted in higher emissions in most cases, except for CO 2. Total CO 2 emissions were

inversely related to speed, the higher the mean speed a vehicle was able to maintain, the less the

total CO 2 was emitted. However, this finding was applicable only to the specific vehicle used in

the experiment and was limited by various testing conditions.

Boriboonsomsin and Barth have conducted a series of research on evaluating the air quality

benefits of HOV Lanes (Boriboonsomsin and Barth, 2007). In 2007, they examined the

operational differences in traffic dynamics between HOV lanes and MF lanes in SouthernCalifornia and calculated the CO 2 emissions and fuel consumptions using the Comprehensive

Modal Emissions Model (CMEM). The results indicated that on congested freeways, vehicles

traveling in HOV lanes produce about 35 percent less CO 2 emissions than those traveling in MF

lanes due to a better flow of traffic in HOV lanes; while on uncongested freeways, although

higher emission and fuel consumption rates are produced on HOV lanes due to higher speeds,

8/13/2019 161142-1

35/85

17

VMT in HOV lanes are much lower, which resulted in a lower emissions mass on a per lane

basis. In 2008, Boriboonsomsin and Bath estimated and compared vehicle emissions contributed

from continuous access HOV lanes and limited access HOV lanes with a combined use of

microscopic traffic model PARAMICS and the microscopic emission model CMEM. It was

found that the limited access HOV lanes contribute to higher amount of emissions due to more

frequent and aggressive acceleration/deceleration maneuvers occurring at the dedicated

ingress/egress sections. The amount of CO 2 emissions were increased by three to eight percent

(Boriboonsomsin and Barth, 2008). In 2009, based on the scientific investigation of the

differences between HOV lanes and MF lanes in terms of their vehicles speed and acceleration

profiles and their fleet composition, HOV lane emission correction factors were developed for

the prevailing speed ranging from 25 to 105 kilometers per hour (kph) to generate emission rates

that are specific for HOV lanes. The analysis results showed that HOV lanes have higher impacton the emission rate of CO 2 as compared to HC and NO X (Boriboonsomsin, 2009).

2.3.2 Bus Exclusive Lane

Guo and Yu investigated the impact of installing different types of bus exclusive lanes on vehicle

emissions (Guo and Yu, 2010). In their study, PEMS data were collected and analyzed to reflect

the vehicles driving and exhaust emission characteristics and traffic simulation model VISSIMwas used to model different design schemes of bus exclusive lanes including central bus lanes,

curb bus lanes, and no bus lanes. This research estimated and compared total CO 2 emissions as

well as emission intensities at different time periods and around interchanges for different design

schemes. Results showed that CO 2 emissions were reduced by about six percent and one percent

after the installation of central bus lanes and curb bus lanes respectively.

2.3.3 Traffic Signal Control Plans

Rakha and Ding (Rakha and Ding, 2003) evaluated the impact of vehicle stops on fuel

consumption and emission rates with the real-world acceleration/deceleration data collected

along a signalized arterial corridor using GPS-equipped vehicles. The study found that vehicle

fuel consumption and emission rates increased considerably as the number of vehicle stops

8/13/2019 161142-1

36/85

18

increased, especially at high cruise speeds. In addition, vehicles fuel consumption was more

sensitive to the cruise speed level than to vehicle stops.

With a combined use of microscopic traffic simulator VISSIM, microscopic emission estimation

model CMEM, and stochastic signal optimization tool VISGAOST, Stevanovic et al. examined a

14-intersection network in Park City, Utah, for the purpose of providing signal timings that

minimize vehicles fuel consumption and CO 2 emissions. Results of the research showed that

after the signal optimization, estimated fuel savings and CO 2 reduction were around one and a

half percent (Stevanovic et al., 2009).

2.3.4 Electronic Toll Collection

Bartin et al. (Bartin et al., 2007) presented a microscopic simulation based estimation of the

spatial-temporal change in air pollution levels as a result of ETC deployment in New Jersey

Turnpike (NJTPK), in which overall impacts, location-based impacts, short-term and long-turn

impacts of ETC system were compared. At each time step of the mobile source GHG emission

simulation, vehicles CO 2 emissions were calculated for each vehicle type based on their speeds

using the macroscopic emission model MOBILE6.2. Results showed that the ETC deployment

reduced overall CO 2 emissions on the network in the short term; however, its long-term benefitswere not sufficient enough to compensate the increase of CO 2 emissions on the mainline due to

the annual traffic growth.

Song et. al (Song, et. al., 2008) analyzed the emissions around a toll station area in Beijng, China,

using PEMS measurements. The real-world vehicle emission and driving activity data for a

light-duty gasoline vehicle were collected simultaneously for both ETC and manual toll

collection (MTC) lanes, and then the emission reductions resulted from ETC lanes were assessed

based on the collected PEMS data. Comparison results showed that CO 2 emissions can be

reduced by 48.9% by using ETC.

Coelho et. al. (Coelho, 2005) studied traffic and emission impacts of toll facilities in urban

corridors. In this research, a methodology that can quantify the traffic performance at a toll

8/13/2019 161142-1

37/85

19

facility with the conventional payment and ETC was developed. This research also explained

the relationship between variables characterizing stop-and-go behavior with environmental and

traffic performance variables, in particular, CO, NO, HC, CO 2, and queue length. The main

conclusion of this work was that the greatest percentage of emissions for a vehicle that stops at a

MTC station was due to its final acceleration back to the cruise speed after leaving the toll

station. When there was a queue of 20 vehicles, vehicles CO 2 emissions can be reduced by 70

percent after the implementation of ETC; while for a queue of only one (1) vehicle, the

corresponding emission reduction was 11percent.

2.3.5 Others

Other traffic management strategies that have been investigated in terms of their impact on

mobile source GHG emissions include demand control, banning heavy duty vehicles (HDVs),

speed restriction, and adaptive cruise control (ACC). In a recent study conducted by Mahmond

(Mahmond et al., 2010), the impact of these traffic control measures on vehicle emissions at a

single intersection located at Bentinckplein in the city of Rotterdam, Netherlands was

investigated with the help of traffic model VISSIM and microscopic emission model EnViVer.

It was found that reducing traffic demand by 20 percent led to about 23percent CO 2 emission

reduction; eliminating the number of HDVs resulted in a total reduction of 25.8 percent for CO 2;speed restriction can reduce CO 2 emissions by 10.7 percent for light duty vehicles (LDVs) and

5.6 percent for HDVs; and ACC reduced CO 2 by 3.5 percent and 1.8 percent for LDVs and

HDVs respectively.

2.4 Summary

The literature review on mobile source GHG emission measurement methods, mobile sourceGHG emission models, and current research on air quality benefit assessment of traffic

management strategies provides solid background knowledge about the accomplishments that

have been achieved in this field and obstacles that exist in the current research, which gives

insightful guide towards conducting the research for this study.

8/13/2019 161142-1

38/85

20

Accurate and abundant data is very important in the research of on-road emissions. With the

ability of measuring second-by-second emissions during the actual use of vehicles in their

regular operations, PEMS is of overwhelming advantage over other existing emission

measurement technologies. Therefore, a PEMS device is used for the data collection in this

research. Considering traffic conditions in Houston, as well as the popularity of traffic

management strategies that have been selected in the existing motor vehicle emission studies,

HOV lane, traffic signal coordination plan, and ETC are chosen as the target traffic management

strategies in this research. In addition, based on the real-world data collected by PEMS, a VSP-

based mobile source GHG emission modeling approach is developed as the basis for emission

estimations. A detailed description about the design of this study is provided in the next chapter.

8/13/2019 161142-1

39/85

21

CHAPTER 3: DESIGN OF THE STUDY

3.1 General Methodology

In this research, the evaluation of traffic management strategies in terms of their effects on

mobile source GHG emissions is fulfilled by a combined use of the field data collected by a

PEMS device and a VSP-based modeling approach. Therefore, the general methodology

includes comprehensive data collection, the application of state-of-the-art vehicle emission

modeling approach, and a thorough evaluation of the selected traffic management strategies.

3.1.1 Data Collection Approach

Two parts of real-world data are needed to perform the proposed evaluation study. One part

includes data collected for the purpose of developing the modeling approach that meets specific

needs of the emission calculation in this study; and the other part includes data collected in the

designed testing areas to facilitate the case-specific traffic management assessment.

The data used for the model development are collected by a light duty gasoline vehicle equipped

with a PEMS unit. The data collection tests are performed on various road types during different

time periods in different days in order to fully capture the testing vehicles typical driving

conditions in the traffic network in Houston.

The data collection for case-specific traffic management assessment is more complicated. It

involves a circumspect design of testing scenarios and a careful consideration of testing vehicles,

testing equipments, testing routes, and testing times and locations. The basic idea is to createdata collection scenarios that are able to reflect real-world traffic conditions with versus without

the implementation of specific traffic management strategies for the purpose of comparison and

then to use real-world vehicle operational and emission data to perform the evaluation study.

8/13/2019 161142-1

40/85

8/13/2019 161142-1

41/85

8/13/2019 161142-1

42/85

24

vehicle speed, engine resolution per minute (RPM), and temperature in second-by-second

resolution and calculating fuel consumption and exhaust flow. An embedded GPS system

provides geographic location information for the testing vehicle at all times (CATI, 2008).

The Axion unit may be placed at a safe and easy to access place in or on the testing vehicle. The

power is drawn from the vehicle power socket or from a cable clamped directly onto the vehicle

battery. For in-vehicle installation, the exhaust sample lines can be routed through a window and

secured to the exhaust system using hose clamps. The operation of the system involves warming

up the equipment before the test, entering correct setup parameters, and checking data for correct

ranges during the on-road test. Figure 1 displays a selection of pictures taken during an Axion-

based emission test.

Axion Installed in the Vehicle Data Collection Operation Interface

Exhaust Sample Line Secured Routing of the Exhaust Sample Linesto the Exhaust System

Figure 1 Selected Pictures during an Axion-based Emission Test

8/13/2019 161142-1

43/85

8/13/2019 161142-1

44/85

26

reported similar values of the exhaust component; (3) review the values of engine operating

parameters, such as RPM, in take air temperature, manifold air pressure, and speed, to make sure

that the data are complete and consistent; (4) check engine RPM to make sure that variations are

reported every few seconds; and (5) compare engine operating parameters with its conventional

minimum and maximum values to determine if there are any irregular data points (CATI, 2008).

A dedicated Macro in Excel is used to eliminate all invalid data.

The most common errors found in the GPS data are data losses and the data deviation caused by

the short term signal block or interference. Therefore, a computer program is developed to

remove deviated data points and smooth the dataset.

3.2.3 Database Establishment

The screened data are processed and then imported into a database developed using Microsoft

ACCESS 2007. This database currently contains three tables: EMISSION table, VEHICLE table,

and DRIVER table. The EMISSION table stores second-by-second test data including testing

date, time, engine RPM, in take air temperature (IAT), manifold air pressure (MAP), gas

analyzer source, fuel consumption, volume and mass of exhaust pollutants, including CO 2, CO,

HC, and NO X, vehicles speed, acceleration, location, bearing, and ambient environmentalcondition, such as temperature, pressure, and relative humidity. The VEHICLE table stores the

testing vehicle information and route information, such as vehicle make, model, engine

placement, year, age, mileage, fuel type, and road types that driving routes cover. The DRIVER

table records the drivers name, age, gender, occupation, years of driving, and contact

information. These three tables are interconnected. For example, if a second-by-second

emission data is located, all the relevant information from other tables can be retrieved according

to the testing date.

3.3 VSP-based GHG Emission Modeling Approach

The basic methodology for developing the VSP-based emission modeling approach is binning

second-by-second VSP data and computing the average emission rate in each bin. The meaning

8/13/2019 161142-1

45/85

27

of each bin is the percentage of corresponding VSP values in the whole distribution, which

captures unique emissions for that emission process. With this partition, the average emission

rate of a particular type of pollutant in that bin for a specified vehicle can be determined.

The accuracy of the VSP-based modeling approach relies on how VSP bins are defined.

However, there has been no clear definition about criteria in selecting VSP cutting points,

therefore, the definition of VSP bins embedded in MOVES is used in the binning process in this

study, since MOVES is currently considered the standard tool that EPA has for estimating GHG

emissions from the transportation sector (EPA, 2010g). On the basis of VSP, speed, and

acceleration, a total of 17 operating modes are defined for running energy consumption for motor

vehicles in MOVES (EPA, 2010h). The definition of MOVES operating mode attributes for

running energy consumption is reorganized into Table 1 according to EPAs guide on theemission rate development for light duty vehicles in MOVES (EPA, 2010i). It shows that, aside

from braking, which is defined in terms of the acceleration alone, and idle, which is defined in

terms of the speed alone, the remaining 15 modes are defined in terms of VSP within broad

speed classes.

8/13/2019 161142-1

46/85

28

Table 1 Definition of MOVES Operating Mode Attributes for Running Energy ConsumptionOperating

ModeOperating Mode

DescriptionVSP

(VSPt, kW/T)VehicleSpeed

(Vt, mi/hr)

VehicleAcceleration(a, mi/hr-sec)

0 Deceleration/

Braking

t a -2.0 OR

( t a < -1.0 AND1t a < -1.0 AND

2t a < -1.0)1 Idle -1.0 t v

8/13/2019 161142-1

47/85

29

3.4.1 Scenario Design for HOV Lane Analysis

There are five major HOV lanes in the Houston area. Based on five consecutive days

observation on the live traffic on those HOV lanes from Houston TranStar, it is found that the

HOV lane on IH-45 North (I-45 N) experiences the biggest traffic demand during afternoon peakhours. Therefore, a freeway segment on I-45 N is selected as the study area. The map of the

study area is shown in Figure 3. In addition, according to Houston TranStar Speed Charts falling

into the selected road segment, shown in Figure 4, the lowest travel speed usually occurs around

5:30 P.M. to 6:30 P.M., therefore, the data collection for HOV lane analysis is conducted during

this time frame.

Under this scenario, CO 2 emission factors obtained on the selected road segment using HOV

lane and mixed flow (MF) lane are compared. Two light duty gasoline vehicles are used for the

data collection, one equipped with a PEMS unit and the other one equipped with a GPS device.

The data collection is performed on two different days. On the first day, the vehicle equipped

with the PEMS unit drives on the designated HOV lane and the vehicle equipped with the GPS

device drives in parallel on the corresponding MF lane, and then two vehicles switch their roles

on the second day of the test. Because the vehicle that runs on the HOV lane requires less travel

time to arrive at the selected exit of the freeway, in order to achieve a better synchronization, it is

designed that the vehicle drives on the MF lane enters the testing road segment about 10 minutes

earlier than does the vehicle that drives on the HOV lane.

8/13/2019 161142-1

48/85

30

Figure 3 Study Area for HOV Lane AnalysisSource: http://www.ridemetro.org/SchedulesMaps/HOV/i45n.aspx

8/13/2019 161142-1

49/85

31

Figure 4 3-Monthly Average Speed on the Selected Freeway Segment on I-45 N(The beginning and ending of the test time periods are indicated by dash lines)

Source: http://traffic.houstontranstar.org/speedcharts/

3.4.2 Scenario Design for Traffic Signal Coordination Analysis

It is observed that the signal timing in Houston Downtown Area is well coordinated. When

driving at about 25-30 miles per hour (mph), the existing traffic signal timing allows vehicles to

pass through several consecutive intersections without being interrupted by red lights. Therefore,

a route composed of four one-way streets in Downtown Houston is selected as the testing area

for this study. As shown in Figure 5, these streets form a closed loop with seven intersections onWebster St. and Pease St. (east and west bound) respectively and four intersections on Crawford

St. and Milam St. (north and south bound) respectively excluding four turning intersections. The

total distance of the testing route is about 2.9 kilometers, in which each pair of adjacent

intersections is approximately 75 meters apart.

8/13/2019 161142-1

50/85

32

Figure 5 Study Area for Signal Coordination Analysis

Data are collected using two light duty gasoline vehicles, one equipped with a PEMS unit and

the other one equipped with a GPS device. These two vehicles take turns running under twotesting scenarios: (1) following the existing coordinated signal timing; and (2) following the

emulated non-coordinated signal timing by making random stops in front of green lights

purposely. The intersection Crawford St. at Pease St. is used as the starting point. Two testing

vehicles always meet at this point at the end of each cycle and leave this point simultaneously at

the beginning of a new cycle. It is designed that both testing vehicles repeat the driving along

the testing route eight times with four times following the existing coordinated signal timing and

four times making random and purposive stops in front of green lights. The type, time, and

location of every stop that testing vehicles made were recorded by a designated person in the

vehicle. This information can facilitate data processing in the result analysis.

Two trial tests are performed before the real test. It is found from the field inspection that all

vehicles on selected streets share the side lanes for through traffic, regulated turning movements,

8/13/2019 161142-1

51/85

33

and roadside parking, therefore, it is feasible to randomly and purposely stop the testing vehicle

in front of green lights without leaving the regular lane, while at the same time without

generating too much disturbance on passing vehicles.

3.4.3 Scenario Design for ETC Analysis

Under this scenario, the amount of CO 2 emissions that a vehicle produces around an ETC station

and a Manual Toll Collection (MTC) station in the same network are compared. After a careful

examination of the type and location of all toll stations in Houston, a segment on Fort Bend

Parkway Toll Road is selected as the testing area. As shown in Figure 6, there are two toll

stations along the selected road segment, the one located in the North accepting both electronic

tolling (pay by a pre-purchased Easy Tag) and manual tolling (pay by coins), which, however, is

always used as an MTC station for this study; and the one located in the South is an ETC station,

which can only be paid by Easy Tag.

Figure 6 Study Area for ETC Analysis Source: https://www.hctra.org/files/Front_SW_Quadrant.pdf

8/13/2019 161142-1

52/85

8/13/2019 161142-1

53/85

35

CHAPTER 4: RESULTS AND DISCUSSIONS

This chapter presents the application of the VSP-based emission modeling approach and the

results generated from the analysis of real-world data collected under each traffic management

scenario. An evaluation of the selected traffic management strategies in terms of their effects on

mobile source GHG emissions is discussed based on the data analysis results.

4.1 Development and Validation of the Proposed Modeling Approach

4.1.1 Data Source

The data utilized for the development of the GHG emission modeling approach were collected

during eight different days from April to July 2010 in Houston, covering different time periods

of the day and various road types. A 1999 Nissan Altima was recruited as the testing vehicle and

was equipped with the PEMS Unit. Detailed information about the testing vehicle is listed in

Table 2. In addition, in order to minimize the influence from individual drivers driving

behavior, the same driver was used for all tests. After the data screening, a total of 35,538 datarecords are left and imported into a database. The whole dataset is then divided into two parts,

one for the model development and another one for the model validation.

Table 2 Information about the Testing Vehicle for Model Development

Make and Model Nissan AltimaEngine Displacement 2.4 LYear 1999Age 11Mileage 71600Fuel Type Gasoline-petrolTransmission 4-speed automaticHorsepower 110.4kw @ 5600 RPMWeight 2925 lbsLength 183.5 in

8/13/2019 161142-1

54/85

36

4.1.2 Development of the Proposed Modeling Approach

A total of 18,253 data records are used for the development of the modeling approach. Based on

the second-by-second speed and acceleration data, VSP values are computed using Equation (2),

and then a computer program is executed to bin these values based on the definition of VSP binsused in MOVES. Figure 7 illustrates the average CO 2 emission rate and data frequency for each

VSP Bin. Table 3 lists the value of the average CO 2 emission rate each bin represents. Because a

vehicles emission rate for a certain exhaust pollutant is mainly determined by the vehicles

technologies, it is a fixed value under different driving conditions. Therefore, the resulted CO 2

emission rates for each bin can be readily used to calculate aggregated mass emissions that each

bin represents in the model application.

Figure 7 Average CO 2 Emission Rate and Data Frequency for Each VSP Bin

0

12

3

4

5

6

0

510

15

20

25

30

0 1 11 12 13 14 15 16 21 22 23 24 25 26 33 35 36

C

Emiso

R

egs

F

e

)

VSP Bins

VSP Distribution

8/13/2019 161142-1

55/85

8/13/2019 161142-1

56/85

38

Table 4 Comparison of Total CO 2 Emissions Based on Modeling Approach and PEMS EmissionData

Bin IDAverage CO 2

Emission Rate

from Modeling

Number of Datain Each Bin

Emission in Each Binfrom Modeling (g)

0 0.8610 1660 1429.23841 0.7480 4468 3341.965711 1.0221 1327 1356.374512 1.3519 1924 2601.063313 2.0718 1158 2399.175714 2.8410 952 2704.674815 3.3750 401 1353.371416 3.3385 129 430.672021 1.2227 823 1006.283722 1.7571 803 1410.957723 2.5174 617 1553.216724 3.0778 521 1603.532225 3.8866 438 1702.333426 4.7534 340 1616.149233 2.7301 363 991.020535 4.1559 707 2938.186036 5.1561 654 3372.0796Total Emissions from Modeling Approach 31810.2949Total Emission from PEMS Emission Data 35173.0460Relative Difference 9.56%

4.2 Evaluation and Analysis of HOV Lane Scenario

According to the scenario design for HOV lane analysis, the data utilized for HOV lane

evaluation are collected between 5:30 P.M. to 6:30 P.M. on July 6 and 12, 2010 on the selected

freeway segment along I-45 N. The same 1999 Nissan Altima is used as the testing vehicle,

which is equipped with the PEMS. On July 6, this testing vehicle drove on the selected freeway

segment using HOV lane, and at the same time, a second vehicle equipped with a GPS device

drove in parallel with the PEMS vehicle on the corresponding MF lane. Two vehicles switched

their roles on the second day of the test. Because the vehicle equipped with the GPS device

always runs in a synchronized way with the vehicle equipped with the PEMS, and only the speed

data collected by the two equipments are involved in the emission estimation based on the VSP-

based modeling approach, the resulting values are comparable.

8/13/2019 161142-1

57/85

39

With a combined use of Google Map and the geographic coordinates recorded in the PEMS unit

and the GPS device, a point close to the HOV lane entrance on the MF lane is marked as the start

point of the target freeway segment; and another point along the exit ramp where both testing

vehicles will pass is selected as the end point. After a data quality assurance procedure, on the

first testing day, a total of 723 PEMS data and 977 GPS data are selected; and on the second

testing day, a total of 707 GPS data and 1,228 PEMS data are selected. Table 5 shows the

resulting CO 2 emission factors on the selected freeway segment using HOV lane and the

corresponding MF lane. It is shown that on the first day, CO 2 emission factors generated by the

testing vehicle on the HOV lane and the corresponding MF lane are 249.55 g/mile and 258.75

g/mile respectively, so the results show that the emission reduction rate by using HOV lane is

about 3.56 percent; on the second day, the resulting CO 2 emission factors on the HOV lane and

the MF lane are 247.15 g/mile and 275.90 g/mile respectively, so about 10.42 percent CO 2 emissions is reduced by using HOV lane.

Table 5 Comparison of CO 2 Emission Factors Using HOV Lane and MF Lane

Total CO 2 Emissions (g)

Total DistanceTraveled (mile)

CO 2 EmissionFactor (g/mile)

Day 1 Day 2 Day 1 Day 2 Day 1 Day 2Using HOV Lane 2797.41 2768.03 11.21 11.20 249.55 247.15Using MF Lane 2882.43 3110.40 11.14 11.27 258.75 275.90Amount of CO 2 Emissions Reduced Using HOV Lane 9.20 28.75

Percentage of CO 2 Emissions Reduced Using HOV Lane 3.56% 10.42%

The difference of emission reductions is mainly due to different traffic demands on the MF lane

in the two testing days. Figure 8 shows the VSP distributions on the HOV lane and the

corresponding MF lane resulted from the two experimental tests. Based on the speed

classification in the VSP definition, Bin 0 and Bin 1 represent the breaking and idling

respectively, Bins 11 to 16 refer to the speed range 0-25 mph, Bins 21 to 26 refer to the speed

range 25-50 mph, and Bins 33, 35, and 36 refer to the speed higher than 50 mph. This figure

reflects that when the vehicle drives on the HOV lane, about 80% data fall into Bins 33, 35, and

36. When the vehicle drives on the corresponding MF lane, the data falling into these three bins

8/13/2019 161142-1

58/85

40

are significantly reduced. On Day 1, about 33 percent of the total data are in Bins 33, 35, and 36,

and only 11 percent of the total data are in Bins 11-16; while on Day 2, the data falling into Bins

11-16 increased to 29 percent, and the data falling into Bins 33, 35, and 36 decreased to 16

percent. Therefore, we can see that the vehicles average speed is lower in the second day of the

test, which also means that the selected segment of the MF lane is more congested on Day 2.

Resulted from the Test on July 6, 2010

Resulted from the Test on July 12, 2010

Figure 8 Comparisons of VSP Distributions on HOV Lane and MF Lane

0%

5%

10%

15%

20%

25%

30%

35%

0 1 11 12 13 14 15 16 21 22 23 24 25 26 33 35 36

V S P

F r e q u e n c y

VSP Bins

VSP Distribution on MF Lane

VSP Distribution on HOV Lane

0%

5%

10%

15%

20%

25%

30%

35%

0 1 11 12 13 14 15 16 21 22 23 24 25 26 33 35 36

V S P F r e q u e n c y

VSP Bins

VSP Distribution on MF Lane

VSP Distribution on HOV Lane

8/13/2019 161142-1

59/85

41

The above comparative analysis demonstrated that during peak periods, the testing vehicle

produces less CO 2 emissions mass per mile by using HOV lane. The more traffic demand on the

corresponding MF lane, the more CO 2 emissions can be reduced by using HOV lane. In addition,

since only a vehicle with a driver plus at least one passenger is allowed to use HOV lane, HOV

lane has the potential to further reduce a vehicles emissions by affecting the total vehicle miles

traveled. Take the results generated from this study as an example: there is one driver and one

passenger in the vehicle running on the HOV lane, if these two persons drove their own

individual vehicles separately on the MF lane during the same testing period, then the amount of

CO2 emissions reduced by using HOV lane can be doubled.

4.3 Evaluation and Analysis of Traffic Signal Coordination Scenario

The data utilized for the evaluation of the signal coordination plan was collected between 3:00

P.M. and 5:00 P.M. on June 17, 2010 in Downtown Houston. The same 1999 Nissan Altima is

equipped with the PEMS. In addition, a 2002 Ford Taurus is equipped with a GPS device, which

runs in parallel with the vehicle installed with the PEMS. In the field test, both vehicles drive

eight cycles around the designed route, with four cycles following the existing coordinated signal

timing and four cycles driving in a manner that makes intentional random stops in front of greenlights. After a careful data review, a total of 5,089 PEMS data and 6,092 GPS data are recorded.

Based on the manually logged time of each testing vehicles exact start and end times of each

cycle and data characteristics, the PEMS and GPS data collected during each cycle are separated

from the whole dataset, and then the amount of CO 2 emissions produced during each driving

cycle is estimated using the VSP-based GHG emission modeling approach.

Figure 9 displays a comparison of the total CO 2 emissions generated by the testing vehicle under

two testing scenarios. When driving under the emulated non-coordinated signal timing, a total of

8472.87 grams CO 2 emissions are produced; while when driving under the existing coordinated

signal timing, a total of 5446.23 grams CO 2 emissions are produced over the same covered

distance. Therefore, the total CO 2 emissions generated under the non-coordinated signal timing

is about 56 percent higher than those generated under the existing coordinated signal timing.

8/13/2019 161142-1

60/85

42

The total CO 2 emissions generated by hour under the two testing scenarios are also computed.

During 3:00 P.M. to 4:00 P.M., the total CO 2 emissions generated by the two testing vehicles are

2617.28 grams per four cycle distance; while that increases to 2828.95 grams per four cycle

distance during 4:00 P.M. to 5:00 P.M. One of the reasons that caused the difference in the

amount of CO 2 emissions on an hour basis is that with the approaching to rush hour, the increase

of traffic flow may compromise the effectiveness of the signal coordination in terms of its

influence on the mobile source GHG emission control. The amount of CO 2 emissions generated

during each driving cycle under the emulated non-coordinated signal timing are mainly subject

to the number of random stops made by the driver, therefore, the total CO 2 emissions generated

by hour under this scenario are not comparable.

Figure 9 Comparison of CO 2 Emissions Following Existing Coordinated Signal Timing andEmulated Non-Coordinated Signal Timing

Figure 10 displays the comparison of CO 2 emissions generated during each driving cycle. It

manifests that when following the existing signal coordination, the average CO 2 emissions

produced by the testing vehicle are approximately 680 grams per cycle distance; while that

increased to approximately 1000 grams per cycle distance when following the emulated non-

2828.954148.79

2617.28

4324.08

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

Total Emissions Following Existing Coordinated Signal Timingotal Emissions Following Emulated Non-coordinat

C O 2 E m

i s s

i o n s

( g )

15:00-16:00 16:00-17:00

8/13/2019 161142-1

61/85

43

coordinated signal timing. Therefore, the existing coordinated signal timing helps reduce about

32 percent of CO 2 emissions per cycle distance comparing with that produced by the emulated

non-coordinated signal timing. It is also observed from Figure 10 that Cycle 2 experienced the

most significant emission reduction, which is about 50 percent. Even though Cycle 3

experienced the least amount of emission reduction, the emission reduction rate still reaches

about 20percent.

Figure 10 Comparison of CO 2 Emissions Following Existing Coordinated Signal Timing andEmulated Non-Coordinated Signal Timing by Cycle

Figure 11 shows the comparison of VSP distributions based on the data collected under the two

testing scenarios cycle by cycle. As we know, a traffic signals level of coordination influences a

vehicles emissions mainly by regulating its stops at intersections and the duration of idling;therefore, the number of data falling into Bin 1 (stands for the operating mode of idling) is an

important indicator to determine how well the vehicle follows the coordinated signal timing.

Figure 11 shows that in Cycle 1, there is the least percentage of data points falling into Bin 1

when following the existing signal coordination. It is also found that in Cycle 2, Bin 1 contains

the highest percentage of data points when following the emulated non-coordinated signal timing.

605.53 614.57734.96

662.22 714.99734.56 723.88

655.52

997.22

1,227.79

910.41

1,188.65

946.55

1,107.33 1,139.39

955.52

0

200

400

600

800

1000

1200

1400

1 2 3 4 5 6 7 8

C O 2 E m

i s s

i o n s

( g )

Cycle Number

Following Existing Coordinated Signal Timing

Following Emulated Non-Coordinated Signal Timing

8/13/2019 161142-1

62/85

44

This result is consistent with the results shown in Figure 10, which demonstrates that during the

testing periods, the trip in Cycle 1 following the existing signal coordination generates the least

CO2 mass emissions, and the trip in Cycle 2 following the emulated non-coordinated signal

timing generates the highest CO 2 mass emissions.

Figure 11 also shows that besides Bin 1, Bin 0 (representing the operating mode of

deceleration/breaking) and Bin 12 (representing the operating mode of cruise or acceleration

with speed less than 25 mph and 0 VSP 3) are also important indicators of the vehicles CO 2

emissions in this study, because Bin 0 and Bin 12 are the bin types that contain a high percentage

of data points next to Bin 1.

8/13/2019 161142-1

63/85

45

Cycle 1 Cycle 2

Cycle 3 Cycle 4

Cycle 5 Cycle 6

Cycle 7 Cycle 8

Figure 11 Comparison of VSP Distributions in Each Cycle

01020304050607080

0 11 1 1 1 1 1 2 2 2 2 2 2 3 3 3

V

Fe

VSP Bins

Existing Coordinated Signal Timing

01020304050607080

0 1 111213141516212223242526333536

V

Fe

VSP Bins

Existing Coordinated Signal Timing

01020304050607080

0 11 1 1 1 1 1 2 2 2 2 2 2 3 3 3

V

Fe

VSP Bins

Existing Coordinated Signal Timing

01020304050607080

0 1 111213141516212223242526333536

V

Fe

VSP Bins

Existing Coordinated Signal Timing

01020304050607080

0 11 1 1 1 1 1 2 2 2 2 2 2 3 3 3

V

F

e

VSP Bins

Existing Coordinated Signal Timing

01020

304050607080

0 1 111213141516212223242526333536

V

Fe

VSP Bins

Existing Coordinated Signal Timing

01020304050607080

0 11 1 1 1 1 1 2 2 2 2 2 2 3 3 3

V

F

e

VSP Bins

Existing Coordinated Signal Timing

0102030

4050607080

0 11 1 1 1 1 1 2 2 2 2 2 2 3 3 3

V

Ds

b

o

VSP Bins

Existing Coordinated Signal Timing

8/13/2019 161142-1

64/85

46

4.4 Evaluation and Analysis of Electronic Toll Collection Scenario

The data utilized for the evaluation of ETC are collected by a light-duty gasoline vehicle

between 3:00 P.M. to 4.40 P.M. on April 12, 2010, on a toll road segment along Fort Bend

Parkway in Houston. Table 6 shows the basic information about the testing vehicle. As theselected ETC and MTC stations are located on the same freeway segment, the testing vehicles

CO2 emissions around these two types of toll stations can be directly collected by the PEMS unit

within the same driving cycle. After the data screening and pre-processing, a total of 5,989 valid

data are recorded for this analysis.

Table 6 Information about the Testing Vehicle for Evaluation on ETC

Make and Model Ford TaurusEngine Displacement 3.0 LYear 2002Age 8Mileage 126,000Fuel Type Gasoline-petrolTransmission 4-speed automaticHorsepower 116kw @ 4900 RPMWeight 3335.6 lbsLength 197.6 in

One of critical steps in comparing the amount of CO 2 emissions produced by the testing vehicle

around the ETC station and the MTC station is locating the corresponding data collected around

the two target places. Based on the experience of toll road driving, it is found that usually a

vehicle starts to decelerate at a distance about 200 meters away from the payment site when the

vehicle is approaching an MTC station; similarly, when the vehicle is done with the payment and

starts to resume the trip, it returns to a comparatively steady speed after 200 meters. Therefore, alength of 200 meters is determined as the threshold to control the length of the roadway segments

that fall into the scope of this analysis. In addition, the maximum queue length observed at the

MTC station in this study is three vehicles. This length of the queue will not make a big

difference on the driving behavior of incoming vehicles 200 meters away from the payment site.

8/13/2019 161142-1

65/85

8/13/2019 161142-1

66/85

48

After the data selection, the numbers of data collected 200 meters around the MTC station and

the ETC station are 814 and 168 respectively. The testing vehicle drives five cycles around the

designed route and passes each type of toll stations two times per cycle, therefore, 10 sets of

emission data are collected around each type of toll stations. Figure 13 shows the comparison of

total CO 2 mass emissions produced within 200 meters around the two stations. It illustrates that

the total CO 2 emissions produced by the vehicle using the ETC station are only 30 percent of

those produced around the MTC station. Figure 14 displays CO 2 emissions produced each time

the vehicle passes through each of the selected toll stations. It is observed that with the

implementation of ETC, emission reduction rates for each individual passing range from 50

percent to 80 percent comparing to using the MTC station.

Figure 13 Comparison of Total CO 2 Emissions Produced around ETC Station and MTC Station

1664.41

4830.35

0

1000

2000

3000

4000

5000

CO2 Produced AroundETC Station (g)

CO2 Produced AroundMTC Station (g)

C

Emiso

g

8/13/2019 161142-1

67/85

49

Figure 14 Comaprison of CO 2 Emissions Each Time the Vehicle Passes ETC Station and MTCStation

It is worth mentioning that during the data collection for this analysis, the testing vehicle did not

spend a long time waiting in the queue and testers who participated in the data collection were

well prepared. In the real-world application, it is possible that drivers may stop longer at the toll

station when paying manually in order to know the correct charges, proper paying methods, andlook for exact coins or cash. All of these actions may result in even higher CO 2 emissions

around the MTC station.

555.20

452.46

562.47

445.90

425.06

455.64

433.28

576.00

488.11

436.23

204.79

216.03

98.13226.46

113.83

175.88134.95

190.68

139.74

163.91

0 100 200 300 400 500 600 700

1

2

3

4

5

6

7

8

9

10

O2 Produced Each Timing Passing the Toll Station (g)

T

mb

oTme

CO2 Around ETC CO2 Around MTC

8/13/2019 161142-1

68/85

50

8/13/2019 161142-1

69/85

51

CHAPTER 5: SUMMARY, CONCLUS